</>

Atul Krishna

Hi, I'm Atul Krishna

Stripped to the essentials. Built to scale.

View My WorkGet In Touch

About Me

Driven by curiosity and powered by engineering precision

I'm a pre-final year Electronics and Computer Science student at VIT Chennai with a deep interest in AI/ML and deep learning. I love working at the intersection of systems, signals, and neural networks, building models that process acoustic waveforms, biometric signals, and satellite imagery.

Alongside my technical work, I'm passionate about community building. As Vice President of Hack Club VIT Chennai, I scale and guide a technical community of developers, organizing hackathons and workshops to foster building culture.

Deep Learning
Web Development
Community Leader
8.65
B.Tech GPA
2,500+
Club Participants
2+
Startup Experiences
5+
Hackathons

Education

Vellore Institute of Technology (VIT Chennai)

Aug 2023 — May 2027

B.Tech in Electronics and Computer Science

Specializing in AI/ML and deep learning. Academic performance: GPA 8.65 / 10.

Navy Children School, Kochi

Graduated Mar 2023

Class X & XII (CBSE)

Class XII: 90% | Class X: 97%. Appointed as Sports Captain.

Professional Experience

Career history and research internships in AI research, startup ecosystems, and tech community leadership

Vice President

Hack Club VIT Chennai

Jan 2025 — Present
Chennai, India
  • Scaled club membership by 40% in 6 months (35 to 50+ students) through structured onboarding and outreach campaigns.
  • Delivered 5+ hackathons and 8+ workshops for 2,500+ cumulative participants, managing logistics, sponsorships, and speaker pipelines.
  • Mentored 20+ junior members on open-source contributions and internship prep, several of whom secured technical roles.

Research Intern

Naval Physical and Oceanographic Laboratory (NPOL), DRDO

May 2026 — Jun 2026
Kochi, India
  • Simulated SONAR target-classification data by generating synthetic acoustic signals at varying SNR (dB) levels to emulate real-world target distances.
  • Converted acoustic signals into Power Spectral Density (PSD) images, reframing the task as 2D image classification instead of 1D signal analysis.
  • Trained a Vision Transformer (ViT) in PyTorch to classify underwater objects from PSD images across multiple categories.

Early Team Member — Fellowship Cohort

CraftHQ

Jun 2025 — Dec 2025
Chennai, India
  • Selected in the top 5% (50 of 1,000+ applicants) for a competitive startup acceleration fellowship with VC access and expert mentorship.
  • Advised 10+ early-stage founders on pitch decks, go-to-market strategy, and investor narratives, contributing to cohort funding outcomes.

Technical Skills

Core competencies, frameworks, and programming tools

Languages

PythonC++JavaCSQLJavaScriptHTML/CSS

AI / ML

PyTorchScikit-learnNumPyPandasMatplotlibMATLABR

Tools & Platforms

GitGitHubVS CodeJupyter NotebookKeilCadence Virtuoso

CS Fundamentals

OOPSDBMSComputer NetworksData Structures & Algorithms

Languages (Human)

English (Fluent)Hindi (Fluent)Malayalam (Native)

Featured Projects

A showcase of production systems and AI models built to solve real-world problems

Freelance Marketplace Platform
Freelance Marketplace Platform
A modern, full-stack Freelance Marketplace Web Application built with Next.js 15, TypeScript, Tailwind CSS, Neon PostgreSQL, and Prisma ORM. The platform connects clients with skilled freelancers, facilitating job postings, proposal submissions, project tracking, role-based dashboards, and reviews.
Next.js 15TypeScriptTailwind CSSNeon PostgreSQLPrisma ORM
DetailsCode
Amazon Product Review NLP Pipeline
Amazon Product Review NLP Pipeline
An end-to-end Natural Language Processing (NLP) pipeline built to analyze Amazon product reviews. Features data ingestion, text preprocessing (noise removal, stemming vs. lemmatization, custom stopword removal), POS-tag based feature extraction, spaCy dependency parsing for feature-sentiment pairing, and high-performance sentiment classification using TF-IDF and Logistic Regression.
PythonNLPspaCyTF-IDFScikit-LearnLogistic Regression
DetailsCode
Land Cover Classification
Land Cover Classification
Trained a U-Net convolutional model for pixel-level land-use classification across 7 categories on 803 satellite images from the DeepGlobe dataset. Engineered an end-to-end PyTorch pipeline with checkpoint resuming, spatial augmentation, and batch normalization, improving val accuracy by ~12% over baseline. Optimized data loading with a custom DataLoader to handle 2448x2448px imagery without memory overflow.
PyTorchU-NetPythonComputer VisionSatellite Imagery
DetailsCode
EchoSafe — Real-Time Public Safety Platform
EchoSafe — Real-Time Public Safety Platform
Architected a real-time distress-detection system streaming live audio/video via WebSockets. Developed a speech- and vision-based distress detection pipeline triggering automated alerts and pushing 10s recordings to a dashboard. Planned deployment on low-cost hardware (e.g., Raspberry Pi-class boards).
WebSocketsSpeech/Audio MLComputer VisionRaspberry PiPython
DetailsCode
Student Mental Health Monitoring System
Student Mental Health Monitoring System
Proposed and built a stress and depression risk-detection system integrating SpO2 readings, meal-skipping, and facial analysis. Fused behavioral signals in an ML classifier to flag at-risk students for early intervention.
Computer VisionML ClassificationBiometric SensorsPython
DetailsCode

Professional Certifications

Industry credentials and expert certifications

OCI Generative AI Professional

Aug 2025

Oracle Cloud Infrastructure

Get In Touch

Ready to bring your ideas to life? Let's discuss your next project and create something extraordinary together.

Let's Connect

I'm always excited to work on new projects and collaborate with innovative teams. Whether you have a specific inquiry or just want to explore possibilities, feel free to reach out.

Location

India
Available for Projects
Currently accepting new freelance projects and full-time opportunities. Expected response time: 24-48 hours.
Send a Message
Fill out the form below and I'll get back to you as soon as possible.